You're a wealth of knowledge on extraction yields, Peppersass. Without merely repeating your insights, suffice it to say I see your pushback on the need for maximum extraction. The only thing I will add, which again Perger has argued well (and with which I agree), is that evenness of extraction may be a more universal goal regardless of road type and other variables. Perhaps there is some reason why someone would want to overextract some of their dose (the fines) and underextract other parts (the boulders), but that reason escapes me.

Wacobipe wrote:This is an intriguing idea as you could toss a lot of coffee before you get to the point of having paid for the grinder. One alternative to consider is the HG-1. While manual, it is a considerably less expensive high end grinder option...but nowhere near as low cost as the sieves.

Thanks for the affirmation and the suggestion on the HG-1. Glad to hear you've had some success with it.

Dpablo wrote:The sieve method seems to me like it won't work. Sieves allow one type of particle size through, and unless it is in the same range as the grind setting, the sieve will effectively be changing the grind setting. Moreover, one is limited in grind setting to whatever the sieve will allow. I suppose you could mess around with dosing to get around this, but it will make things much harder.

I'm not sure I follow, Dpablo. The reasoning behind this [hypothetical, tentative] method is precisely for the reasons you mention. Yes, it would effectively change (or, more refine) the grind setting. Maybe there's some other point I'm missing.

TomC wrote:I've tried this before and it just makes a mess and simply doesn't work. At espresso fineness, there's a lot of static clinging and cohesive forces that make it difficult to separate. A good portion of fines are practically magnetized to other smaller particles. You'd have to set up a Ro-Tap, most likely, to get a decent yield, while still wasting a ton of coffee, making a bigger mess, etc.

Okay, I'm really interested to hear more about this, TomC. It sounds like you worked with a laboratory test sieve, which is what pretty much everyone has been using up until recently. Do you remember what micron sizes you were using? The makers of the Rafino have mentioned that anything smaller than 200 microns starts getting futile for the reasons you mention. And yes, I've actually considered a laboratory sieve shaker to increase yield, but they're prohibitively expensive, not to mention bulky and would take a ton of time just to pull a shot.

coffeekid wrote:[Long-time reader, first-time poster here. Be gentle.

Hi Peter, and welcome to H-B. Great first post.

coffeekid wrote:This is much less scientific, but in the end if feels like there is a lot of dubious and motivated reasoning in support of the particle distribution of high-end grinders.

Let's be fair: your post shows obvious biases and inconsistencies. First you put the EK43 on a pedestal as the holy grail of grinders, using terms like "delicious", "amazing", "fantastic". Then you proceed to trash high-end grinders in general. You can't have it both ways.

coffeekid wrote:All that said, there is one other concern I've heard floating around, which I have no way to verify or falsify: Espresso brewing NEEDS a wide range of particle sizes in order to maintain the structure of the puck in the brewing process. ... Though I have no way of disputing this claim, this seems to be a convenient fallback in support of high-end grinders.

You DO have a way of disputing this claim: design and perform experiments to (dis)prove this hypothesis. But my particle sizing experiments, experience, and common sense all lead me to believe this claim is correct. A heterogeneous mixture of particle sizes will pack more tightly, and help control flow through the espresso puck. For non-espresso brewing, this is much less of an issue.

TomC wrote:I've tried this before and it just makes a mess and simply doesn't work. At espresso fineness, there's a lot of static clinging and cohesive forces that make it difficult to separate.

+1. A finer grind setting will always produce more fines. Sieving is probably better at eliminating boulders than dust. The good news about fines: although numerous, they represent only a small fraction of an espresso grind by weight (mass), and are unlikely to have a major impact on taste.

My question would be: How much manipulation of the grounds would yield the desired results and would that amount of handling degrade flavor? My question at this point revolves around the time it would take to separate the "unwanted" grounds, and exposure of the grounds to the air.

Additionally, waste was mentioned, and in combination to roast levels and origins, how much additional waste would be created when searching for the "correct" set if sieves for any given coffee?

Tom C's mention of static and how that would affect the results was very interesting in terms of the "magnetic cohesive" nature of the grounds. Using an older Baratza Virtuoso to grind for Turkish. I have had a good amount of ground literally jump out of the catch cup when removing it from the grinder.

Scientifically, it is an interesting hypothesis, in terms of practically, by the time you came up with solutions to the issues described in this thread, the solutions could end up costing more than a good grinder.

Hi Peter,

An interesting topic - I am in the same boat(pre-ordered Rafino as well) and in the same thought camp.

However, having tried EK43 shots, I don't think it is what it all claimed to be. EK43 is probably so mystified because it's hard to find/buy, and it's hyped(personal opinion of course) at the moment. I've tried EK43 shots, to be honest, it's delicious alright, but not what I seek in an espresso.

Matt Perger is very generous at sharing information (kudos to him and I do follow his postings). But he tends to plaster hypothesis (but not hard-proven yet) to explain what he's seeing. Nothing's wrong with that approach and it is what I would do. The result of that is misinterpretation can happen when the hypothesis is directly applied onto other context.

For instance, about EK43 being the holy grail. Not true to me - at least for my personal preference. I might be wrong but EK shots seem like an 'espresso' made for brewed coffee drinkers (I don't enjoy as much nor care much about drinking brewed coffee)?

About evenness in extraction, some of the latest posts were saying that it's physically impossible to extract evenly. True if you think about it - the particle sizes are varied even on EK, and there's the temperature gradient in puck. So the claim about 'even' extraction being better is not proven at the moment.

Put another way, if we want a more complex taste profile, does even extraction help us in that direction?

About EY%, I agree with Pepperass(Dick)'s view. Absolute number doesn't mean much by itself and higher isn't better. EK's sweet spot is at 24%, conical's at 20% and flat is at 22% (very very broad generalization here). What all those numbers are telling us is that they taste most balance in that EY% region. Conical with its terrible distribution doesn't neccesarily taste worse than EK's 24%. The 20% EY of conical most likely contains different flavor molecules(including acid and bitter components) in different ratio than the EK's 24%. Plus, all EY measurement is disregarding undissolved solids and any other components that aren't water soluble (which can contribute to taste experience). So, we just can't compare them based on number alone.

I've been making some single basket brews a lot lately, and they are mostly at 13-14% EY (seems to be the ceiling EY and not grinder related). However, the taste is so delicious, so much more than the higher EY% from a VST double basket. Another case for why higher EY isn't always better.

Another caveat for sieving (and laser diffraction particle analysis) - both methods can't tell us what the shape/morphology is like. The result of that is you can have 15g of 200.0 micron particles, but will still extract differently due to different grind surface/shape. So for this reason, there is a possibility you can't replicate EK43's grind profile even with sieving.

Another uncertainty about sieving is shot-to-shot consistency - will sieving allow repeatable pull and dialing in properly? As mentioned, the greatest improvement from sieving would be to remove the boulders - that is if the unimodal theory holds true. I am keen to see the result when my Rafino arrives. In the mean while, let's keep the discussion going.

Forgive me if someone already mentioned this, but in addition to consistency in particle size, consistency in particle SHAPE is likewise important. My understanding is that higher quality grinders also deliver a consistency in particle shape that lower quality grinders cannot match. Shape consistency is a variable that using a sieve would not impact.

EDIT: Was on my phone when I posted and didn't see the post above mine

Thanks for the sustained engagement on this issue, John and Randy! I've benefited a lot from your prior posts, and I'm a huge fan of the WDT (John), so I appreciate your many contributions.

There's no question I have biases in this discussion (who doesn't?), but I can honestly say that I have no allegiance to the EK43 nor do I aim to bash any particular setup-high-end or otherwise. I offer the EK43 as an example primarily because most baristas competing at WBC 2016 were using it and Matt Perger and others have recently begun praising it highly for its potential with espresso (perhaps for the reasons mentioned by Peppersass regarding the trend toward lighter roasts). The only defense I'd offer is that the inconsistencies you rightly note might just be ambivalent suspicions about just how sound the justifications for machines like them are.

[Personal note, which is mostly irrelevant to this discussion: The truth is that if I can understand the analytical argument for why a machine like the EK43 (or Robur or Super Jolly, it matters little in my mind) is better than my hypothetical sieved scenario, I will most likely start saving for something like it. You're of course correct, John, that there are obvious ways of experimenting with this hypothesis. Regrettably, at this moment I just don't have the money or equipment (old, essentially useless espresso machine; mediocre grinder; Rafino sieve system on order), so at least for the time being this conversation remains theoretical for me. [I've actually volunteered at some community coffee shops just to play around with better equipment. Exploitation, I know.] As you might imagine, the "question behind the question" here is about future grinder investments.]

Regarding the waste and feasibility of sieving fines, and espresso in general: First, I have no problem conceding that magnetic cohesion, static, and clogging are issues. At least with current equipment, there are serious obstacles to overcome in order to make sieving convenient and precise. Also, Randy, the issue of oxidation from excessive handling and sieving is an important variable that, frankly, I don't have the technical knowledge to think through. Suffice it to say, I'm not so committed that I'd by sieving in a vacuum.

Yet I am not persuaded by the argument that eliminating undesirable grind sizes from a dose is wasteful. In fact, it makes far more sense that the opposite is the case. Even if one loses something drastic in the sieving process (e.g. - 25% of one's dose) AND one for some reason decides not to repurpose that 25% (again, I'm a huge fan of Turkish), if that lost 25% has a negative effect on the extraction and requires that one adjust one's other variables (time, temp, pressure) to get the "optimal" extraction from a mix of 75% good and 25% bad grind sizes, then NOT sieving would actually waste more of the brewing potential of that dose. To make a crude but fitting analogy: No one argues to keep dirt, rocks, and twigs in sacks of coffee just because doing so reduces the total weight of the harvest, so (other than for matters of convenience) why argue to keep in undesirable grind sizes in one's dose?

I continue to be perplexed (genuinely, don't misread this as sarcasm) by the argument in favor of a bimodal distribution, though I'll likely have to defer to your experiences, John. On the matter of water flow, it's hard for me to understand how a homogeneous mixture would somehow halt all water flow. To borrow from a weak metaphor that apparently has something to do with life priorities, it would seem to me that the fines would fill in the voids of the larger modes, reducing water flow. And, if that restriction is what is desired, again it would seem to make more sense to simply make a unimodal distribution smaller under the desired water flow restriction was met. I'll admit that my theoretical approach risks being unhelpful for practical situations. At the same time, I feel that thinking through these hypothetical situations can help with future "backwards design" for grinding and brewing methodology.

I realize this is very theoretical, but consider the following two scenarios. Let's say someone hopes to yield 40g of brewed espresso from a 20g dose at 201 degrees Fahrenheit and nine bars of pressure. Let's even say that (for some reason) they want this extraction to last 35 seconds. And let's say that they are able to achieve this with their current, unsieved bimodal dose, which has the following bimodal distribution:

Scenario one: 75% or 15g of dose is exactly 300 microns; 25% or 5g is exactly 100 microns; mean particle size is therefore 250 microns.

Again, I realize that it's unrealistic or impossible to achieve this perfect bimodal distribution. But, in this scenario, the best one can do is target the parameters of one's brew at 250 microns, which is a cruel irony because, in fact, none of the grind is actually that size. The effect would be that the fines (100 microns) are overextracted and the semi-boulders (300 microns) are underextracted. [I know, this is common knowledge] So, technically speaking: Yes, none of the dose is being wasted. But at the same time, none of the dose is being properly extracted.

In the second scenario, the siever (that is, the barista who sieves) effective decides, "Rather than improperly extracting 100% of the dose, I'll try to properly extract 75% of the dose." They adjust their bimodal grinder so that the upper mode is at 250 microns or until the other brewing metrics (i.e. - yield, time, pressure) are achieved. This will also have some effect on the finer mode, perhaps making them even finer, though not necessarily:

Scenario two: 75% or 15g of dose is exactly 250 microns; 25% or 5g is exactly 80 microns, but is removed. Total dose is now 15g, but is entirely 250 microns.

Without belaboring the obvious, this second scenario makes the most of the vast majority of the dose, and can truly tailor their brew parameters to their actual particle distribution. 25% is "wasted," but not missed.

I'll admit that my theoretical approach risks being unhelpful for practical situations. Achieving these levels of perfect modality in either scenario would be difficult if not impossible. At the same time, I feel that thinking through these hypothetical situations can help with future "backwards design" for grinding and brewing methodology. In short (and to bring it back to the OP), I'm trying to understand some of the principles that are being used to justify certain grind distributions.

coffeekid wrote:On the matter of water flow, it's hard for me to understand how a homogeneous mixture would somehow halt all water flow. ... it would seem to me that the fines would fill in the voids of the larger modes, reducing water flow. And, if that restriction is what is desired, again it would seem to make more sense to simply make a unimodal distribution smaller under the desired water flow restriction was met.

Nobody is claiming a unimodal distribution will halt flow (quite the contrary). And yes, by grinding finer to compensate for sieving fines, you should be able to reduce the flow rate of a unimodal distribution.

<speculation>
But based on Jim Schulman's experiments with a large flat burr grain mill, you may find it difficult to dial in a good espresso from a unimodal distribution. His punchline: "I haven't been able to get any sort of consistent espresso extraction from this grinder. The fines are just too low."
</speculation>

Please don't take this as an indictment of sieving. Give it a try and let us know what you find out.

I have a 300 micron sieve sitting right by me at this moment. It is 20 cm in diameter. This size of sieve will allow most of the fines to go through and not much else. I have sieved freshly ground coffee many times. If I start with 20 grams of freshly ground coffee, best results to me are obtained by shaking the sieve until 5 grams of fines are obtained. The fines are wonderful for Turkish!!!! The remaining 15 grams makes good espresso. I have done this with a Pharos, Hg-one, and Rocky. (I no longer have the Rocky.)

With the fines missing, one needs to grind quite a bit finer to obtain a normal shot. To me the flavour is decidedly different. As one would expect, the "chromatography" effect is greater with a more uniform size of particles. Standard laboratory chromatography column packing is produced with very uniform sized particles.

[Wow, loving this discussion, for the record.]

I really appreciate your hands-on reflections regarding the EK43, Sam. I agree that there seems to be a ton of hype, and I'll confess that (despite my healthy suspicion in this thread) I've perhaps fallen into the trap of obsessively anticipating what the EK43 Barista will end up looking like and performing. I also agree that the matter of particle shape (also raised by Daniel) is incredibly important and perhaps unmanageable by sieves. In truth, that's another variable that I'm simply unequipped to analyze well-at least yet.

Though (not the be the perennial contrarian), I'm not sure I follow the arguments for "complex > evenness." The only definition of "complexity" I can understand vis-à-vis yield is that there is a combination of underextracted and overextracted coffee. So, one could have an "even" 18% extraction (in which all or most of the coffee is equally extracted) as well as a "complex" 18% extraction (in which there is both overextraction and underextraction). I know I risk sounding simplistic, but it's hard for me see "complex" as anything but a surrogate for "imbalanced extraction." [Again, I realize that this "perfectly even" extraction I'm describing is probably an impossible ideal. Even if "complexity" were a veil for overextraction, it's not like we can avoid it entirely.]

Also, to be clear, I also agree with you, Sam, that there's no getting around matters of consistency. So, while my interest in sieving is motivated by frugality, there's no question in my mind that sieving might actually work BEST with high quality grinders. That is, with a great unimodal grinder's narrow and consistent distribution, it would make sense that sieving would only IMPROVE an already great grinder's consistency and quality of extraction.

RapidCoffee wrote:Nobody is claiming a unimodal distribution will halt flow (quite the contrary). And yes, by grinding finer to compensate for sieving fines, you should be able to reduce the flow rate of a unimodal distribution.

Thanks for clarifying the matter of unimodal distribution and its relation to water flow, John. Perhaps I've misunderstood the arguments of others (not you), but it does seem that some people argue that it is only because of the presence of fines that good waterflow is possible.

Robert: It's exciting to hear from someone actually using this method! Your method of deciding to stop sieving once you've separated a certain mass of fines (presumably, you're sieving fines into a bowl on a scale?) is the sort of practical compromise I feel would be necessary to make this work. Though it makes sense in theory, I'm glad to hear you've in fact experienced this need to reduce the modal grind size in order to obtain a normal shot. Have you tried any other sieve sizes? Even though you've had some success, I do fear that sieving might become exponentially more difficult when jumping to 250 and 200 micron sieves, which might be important for making the method work well for espresso.

I swear friends, send me $5k of equipment and I'll do an obsessive write up investigating this method.

Interesting publication relevant to the thread. http://socraticcoffee.com/2016/06/explo ... xtraction/